### Noureldin's blog

By Noureldin, history, 2 weeks ago, ,

things to do in quarantine

• train to be better at CP
• get some work done
• generate some CF statistics

#### vanishing rating points

so I ran a python script that collected all rating changes of every rated round and I found some interesting stuff which can be found at the end of blog, however the thing that caught my attention is that rating points are vanishing at a rate of ~5.483 points/round/contestant!!!

the average rating change is negative ~ -5.483 points/contestant/round which is weired, you would expect that number to be around zero since the rating system scales everything so that the sum of ratings of contestants after each round remains the same as it was before the round, maybe a side effect of rounding ?

to verify this I compared the sum of rating of all rated accounts against what it should be, there are 258434 accounts that competed in at least one rated round, the sum of current rating of those accounts is 364795845 which is 22855155 (5.895%) less than what it should be 1500*(number of accounts) = 387651000

so why no one notices and ratings generally increase, it's -I think- because of the large influx of new accounts, each new account adds 1500 points to the pool of points once it starts participating in rated rounds so no one will notice ~5.483 points lost per contestant per round

if you think of the points as money and rating changes as transactions then you most likely won't notice that 5$vanishes with each transaction from your account if each transaction involve$1000s, so basically we are acquiring points/money a lot faster than we are losing them and we won't start to feel the loss until the number of active accounts saturates, unless I'm missing something?!

so in a hypothetical scenario in which no new accounts are created and we just keep competing against each other and assuming all accounts are active we will all have ~ zero rating after ~257.44 rounds :V (or 6653.22 rounds if the number of contestants in each round is the -currently- more typical number of contestants of 10K)

#### how long you will need to reach high rating

you -probably- need -if ever- around 2.36 the time you needed to become a master to become LGM

the average number of rounds that contestants needed to hit a rank for the first time is summarized in this table

rank average number of rounds to become
Master 27.163
International Master 34.818
Grandmaster 39.651
International Grandmaster 47.638
Legandary Grandmaster 64.038
rank/percentile 0th25th50th75th100th
Master2.0 8.0 18.0 38.0 250.0
International Master3.0 12.0 25.0 48.0 213.0
Grandmaster3.0 15.0 30.0 55.0 274.0
International Grandmaster5.0 21.0 39.0 66.25 198.0
Legendary Grandmaster8.0 31.25 54.0 88.75 221.0
I wanted to list the accounts that reached each rank in minimum number of rounds however I found that most of them are fake accounts :\ :(

• +67

By Noureldin, history, 3 months ago, ,

UPDATE: it's back ^_^

so it came to my attention a few days ago by fegla that icpcarchive only gives WA or CE verdicts (judging status) to make sure I submitted a couple of my already AC codes and they gave WA ... so it seems that it's not functioning properly.

I don't know if it's down for good or temporarily but livearchive (though admittedly not the nicest of OJs, I mean who didn't get frustrated because of WA due to \n or a space or a weird I/O format or corrupted testing data) was a great repository of regional CPC and ICPC contests (with contests dating from 1988 from all over the world) and it's really sad to lose it.

so is there anything that can or is being done to save it -and while at it improve it- ?

• +13

By Noureldin, history, 7 months ago, ,

since ACPC is drawing nearer and most regional contests are over, I'm starting a list of teams attending ACPC.

please comment with the name of your team name, university, country and team members' accounts or of teams you know.

UPDATE: the ACPC will be held between Nov, 29th and Dec 3rd in Sharm Elsheikh, Egypt.

Country University Team Name Member 1 Member 2 Member 3 team rating
Jordan PSUT Aurora Motarack Hiasat Dark 2561
Egypt unofficial high school team Room 116 mohammedehab2002 Osama_Alkhodairy zoooma13 2388
Egypt AAST (unofficial) Too late :( Yousef_Salama Omar_Elawady tamp_ 2345
Egypt FCIS Ain Shams 3 x 1 Abdelrhman_Akram islammohsen Bohoty 2314
Egypt FCIS Ain Shams Compilation Error missing 'slifer' compiler_101 Error_404 XsliferX 2307
Egypt AAST EnthusiasmOverflow MahmoudBadawy Noureldin Yousry 2299
Syria Damascus University Assert the.last.wizard Mouaz_Alkhodari EsmaelSY 2290
Tunisia INSAT NP-BOSTAJI lionadis letmeC svanO 2254
Syria Damascus University How you doin?;) Mohammad_Kartoumeh Turing.Machine khaled97ha 2224
Egypt Faculty of Engineering Alexandria university IGN Mohammed_Badawi Ahmed_AIansary Evro 2199
Egypt FCI Cairo University Be Done OsamaFathy alimagde omar22 2188
Jordan unofficial high school team JOI3310 black_ghost omarnaseer adhami 2165
Syria Tishreen University Too lazy to propagate NourAlhadi engineer_z C137 2047
Lebanon American University of Beirut Segment Team ChaosAngel Charbel11 0-jij-0 2036
Syria Al-Baath university Peaky Blinders Mll Akheer SyrianTony _Mugiwara_ Marcel_Ib 2025
Egypt Assiut University Ya 5rabyy MuhammadMoustafa _AhmedHafez Kolo_Ray7 2015
Egypt Shoubra FE-BU BugLife IsmailToukhy Elwazer _Samir 2006
Egypt Faculty of engineering Cairo university WF ISA Half-Blood_Prince AbdallahHemdan Hassanosama 2005
Egypt unofficial high school team glad to take your input MohamedAhmed04 Mohamed.Sobhy --PATH-- 2001
Egypt GUC Educated Noodles YahiaSherif Hemose O_E 1995
Egypt Assiut University Ala7lam_team Mohammed_Atef_Hassan AhmedEzzatG karemo 1986
Egypt FCIS Ain Shams While(++Procrastinators) Khaled-Mohamed _HossamYehia_ HazemEssawi 1983
Syria Al-Baath university Can we defeat SyrianTony? theknife2001 mourad_tu Naseem17 1980
Egypt FCI Cairo University Stack Overshab7na KingSlayer. Dardy MonaA 1958
Syria Tishreen University Acceptophillia MahmoudAlio math_master mr.spring 1951
Syria Damascus University ICPCNextCentury SaeedSabbagh abdalkader-s Bsher 1931
Tunisia Tekup university IDERSHIP BigDaddyPanda _Send_Nodes_ Darnze 1877
Egypt Faculty of Engineering Alexandria university Finding Dory AbdulRahman_Nour walaa.khafagy LouaiZahran 1837
Egypt Tanta University -3iq abdo_taha fawzy20 zaher_ 1812
Tunisia Tunisia Polytechnic School GG&GG/Yes __monta Haykoul Rm6 1811
Egypt FCI Kafrelsheikh University No pain no balloon YahiaAshraf74 It_Wasnt_Me Eltoney 1780
Palestine Palestine Polytechnic University 7Up ItsMeMario Logic_PS Mr-Spy 1779
Egypt Suez Canal University Mal2yna4 esm ll team m4 7war y3nee KerollosNabil khaledrezk kerolloz 1765
Egypt Faculty of engineering Cairo university MEKASO Muhammad98 kareem98 GO_oGO_o 1744
Egypt Faculty of Science — Cairo university Faree2.L.Batt Abdullah_Alabd One-Eyed.OWL moazamar310 1735
Egypt Minia University java++ abdelrahman_mohammed GM.AbdelrahmanWael A-1_1-A 1683
Oman Sultan Qaboos University Null satillendlim n3sser massar 1678
Oman Sultan Qaboos University Syntax Error Ahmed_Allawati mashal1996 khalfan 1674
Syria Al-Baath university Dark mode Mohamed-Nour yasso001 omar94 1670
Palestine Birzeit University Limitless laithmubarak11 Masalmah98 y.shqair 1651
Egypt unofficial middle school team BLE ahmedfouadnew Yasseenkamel Aihay 1602
Jordan Al- Balqa' Applied University (BAU) Teeny Tiny Posibillity Batool_AbuRumman Moayd_abual-rub Abukhadijah 1598
Egypt Nile university Brows Menna-Allah05 Hossam._.Ali Khaled.Emad 1562
Egypt Helwan University ayasm abo_eldahab L-Lawliet waggia 1480

• +79

By Noureldin, history, 23 months ago, ,

Hello CodeForces! I'd like to invite you to the online mirror of the 2017 ACM egyptian collegiate programming contest ,which will be held on Wednessday, 6 July 2018, 02:00pm Cairo time

ACM ECPC 2017 Highlights:

the problems were prepared and tested by : Thrax -chief judge-,Badry, Noureldin, SyrianPetr, safrout, Lvitsa, amrSamir, Amirnasr, islam-al-aarag, justHusam and RedStone

problems authors:

Assessment — islam-al-aarag

Cheering parabolas — Thrax

Dream Team — Lvitsa and Thrax

Evaluations — Noureldin

Forgot the Flag! — Noureldin

Half Nice Years — Lvitsa and Thrax

Important matches — justHusam

Jacking ticket — Thrax

Katryoshka — Lvitsa

Lazy ERCD — Mohamed Fouad 

You will be given 5 hours to solve the problems.

Good Luck and have fun!

UPDATE as some of you found out the test sets weren't complete and only samples were ported to the gym version ... that's due to a mistake I made while porting the contest

what happened?

I tried importing the contest normally by coping the contest.xml to the ftp server but it generated alot of errors which I couldn't understand at the time so I imported the problems manually but forgot to run the doall.sh and as a result none of the big test files were generated

now the contest is part of 'codeforces history' so I can't modify it, so I created another one with the original tests used in the contest so have fun with ECPC17

• +103

By Noureldin, 5 years ago, ,

for a competitive programmer the most important features are what C++11 has already introduced namely (range based loops ,lambda expressions ,brace initialization of aggregates, unordered containers ,..... etc) which can be found here .

1) numeric literals

now you can separate digits of a number with single quotes ( ' ) to improve readability

auto num = 10'000'000; // num = 10^7

this works for floating numbers as well

auto fnum = 0.000'015'3; // fnum = 1.53 * 10^-5

now you can initialize numbers using their binary form this feature was experimentally included in C++0x (the prototype of C++11) however it didn't become official until the announcement of the new standard

auto bnum = 0b0100'1100'0110; // bnum = (010011000110)2 = (4C6)16 = 1222

//note that the spacing of the quotes doesn't matter

auto snum = 1'0'0'000'00; // snum = 10^7

2) std::gets was removed due to its tendency to cause RTE due to overflow in its buffer

3) lambda expressions can now be generic:

in C++11 you had to specify the type of the inputs to the lambda expression

auto lambda = [](int x, int y) {return x + y;}; // C++11 --- had to specify type of x and y
auto lambda = [](auto x, auto y) {return x + y;}; // C++14 the compiler will decide the type for you ;)


4) improved std::tuple (in particular improved std::get)

in std::tuple introduced by C++11 (which was experimentally added to TR1 in 2007) to get the value of an element you had to use only its index

int element = std::get<2>(mytuple); //get the third element in mytuple

now you can do the same thing plus the ability to index by type

std::string element = std::get<string>(mytuple);// return the string element in mytuple

note: if the tuple had more than one element of the same type using std::get that way would cause compilation error

a feature of C++11 that's not mentioned in the link above that I find very useful

C++11 introduced the ability to pass rvalues by reference that's to pass temporary values by reference such as

// compilation error in C++03 .... legit in C++11 & C++14
string mul(string && a,string && b);
mul("10","30");
`

as you can see from the specific example I used ,it speeds up calculuations in particular this snippet is from my biginteger implementation which sped up biginteger multiplication by 30-40% on average (tested on numbers containing ~ 50000 digits)

note: some legit C++03 && C++11 codes might break if compiled using C++14 standards

codes that use std::gets are the simplest case but there are others for more details you can refer to this link

finally ,there are some other additions and modifications in C++14 that are very important (relaxed constexpr restrictions ,improved lambda capture expressions ,the attribute [[deprecated]] ,heterogeneous lookup ,....etc) to name a few, yet their importance to a competitive programmer is less than that of the mentioned features (at least from my point of view)